A stepwise approach using metabolic volume and SUVmax to differentiate malignancy and dysplasia from benign colonic...

The purpose of this study was to suggest a new diagnostic strategy using metabolic volume (MV) and maximum standardized uptake value (SUVmax) to differentiate malignancy and dysplasia from benign colonic 2-deoxy-2-18F-fluoro-D-glucose (FDG) uptakes.

From records of 21,317 consecutive FDG positron emission tomography/computed tomography (PET/CT) scans at 2 centers, 102 focal colonic lesions in 99 patients investigated by colonoscopy and histopathologic examination were eligible for this retrospective study. SUVmax and MV were compared according to colonoscopic and histopathologic results. Firstly, dysplasia was separated from malignancy and benign lesions. Secondly, malignancy and benign lesions were separated from each other. The better parameters of each step were determined, and a diagnostic strategy was developed from their combination.

A total of 102 incidental colonic FDG uptakes were revealed as 32 malignancies, 43 dysplasias, and 27 benign lesions. MV better differentiated dysplasia from malignancy and benign lesions (cutoff value, ≤3.14 cm3; area under the receiver-operating characteristic curve [AUC] = 0.947), and SUVmax better differentiated malignancy from benign lesions (cutoff value, >9.1; AUC = 0.934). Overall, the stepwise algorithm using MV and SUVmax (AUC = 0.886) was superior to single measurements of SUVmax (AUC = 0.750) and MV (AUC = 0.714) for differentiating malignancy and dysplasia from benign lesions; sensitivity: 92%, specificity: 85%, accuracy: 90%, positive predictive value: 94%, negative predictive value: 79%.

The stepwise approach using MV and SUVmax was able to differentiate malignancy and dysplasia from benign colonic FDG uptakes on PET/CT. Colonic FDG uptake with MV ≤3.14 cm3 had a high probability of dysplasia. MV >3.14 cm3 and SUVmax >9.1 indicated malignancy, whereas MV >3.14 cm3 and SUVmax ≤9.1 indicated benign lesions.